PROGRESSION OF ACUTE KIDNEY INJURY TO CHRONIC KIDNEY DISEASE: CLINICAL AND EXPERIMENTAL INSIGHTS AND QUERIES

Abstract

There are an increasing number of clinical suggestions that acute kidney injury (AKI) can be complicated by the onset of progressive renal disease. Indeed, given the frequency of AKI in hospitalized patients, were AKI to initiate disease progression, it could potentially be a leading cause of, or contributor to, end stage renal disease. Insights into the natural history of AKI, and potential mechanisms for disease progression, can be gleaned for experimental studies. Although such studies underscore the principle that AKI can “heal with defects”, whether ongoing renal disease develops remains a subject of debate. Indeed, in the aftermath of AKI, a variety of secondary renal protective pathways are activated which may retard or prevent severe, chronic kidney disease (CKD). Furthermore, the onset of acute uremia, per se, may exert surprisingly potent renal protective effects. The purpose of this brief report is to review some of the clinical and experimental data that deal with these complex issues.

Introduction

Although severe AKI has existed throughout history, it was not until the introduction of acute dialysis ~50 years ago that AKI’s natural course could be defined. This was made possible because dialysis allows for sufficient time for renal recovery to occur and for possible subsequent disease progression to develop. Early investigators reported that “the vast majority of patients who survive achieve clinically normal renal function despite frequent subclinical and histological defects” (reviewed in ref. 1). These included 20–40% reductions in GFR, urinary concentration and acidification defects, and histologic evidence of interstitial fibrosis. It was also suggested that “in a small number (<5%) of patients in whom renal recovery is incomplete, a secondary regression of renal function may be observed” (1).

During the past 10 years, there has been increasing interest in, and awareness of, the prospect that AKI can indeed progress to CKD. However, it still remains uncertain as to the frequency of this transition and its underlying mechanisms. One critical study which addressed this issue came from Lo et al (2). These investigators followed the course of 703 patients with AKI which required acute renal replacement therapy. Of these individuals, 508 survived, allowing for their subsequent renal course to be discerned. Based on that study, it was suggested that there was a 28 fold increased risk of developing either stage 4 or stage 5 renal disease (2). However, it would appear that, in the vast majority of cases, this outcome was apparent within 3 months of the onset of AKI. Indeed, beyond that time point, the Kaplan Meier slope for renal disease onset or progression was essentially identical to that observed in a non AKI cohort of hospitalized controls (2). Thus, it remains unclear as to how many patients actually developed progressive disease above and beyond the early renal dysfunction / “healing with a defect”, as noted by early investigators of this field (e.g. ref. 1)

Can experimental AKI models provide insights into the course of AKI and possible disease progression?

In order to better define the natural history of AKI, investigators have turned to experimental models. However, an inherent limitation of these studies is that the severity of the initial renal insult must be severe enough to induce substantial tubular damage, but yet mild enough to prevent early animal death from acute uremia. This is because of the technical difficulties inherent to supporting experimental animals with dialytic therapy. Despite this limitation, important insights have been gleaned. In this regard, seminal studies by Basile et al have employed the bilateral renal ischemia model to explore these issues (3, 4). One such study employed a rat model of 60 min of bilateral renal ischemia, with subsequent renal functional and histologic assessments performed over the ensuing 40 weeks (3). Noteworthy in this regard is that 40 weeks represents ~25% of a normal rat life span, clearly long enough to ascertain AKI’s natural course. In those studies, severe initial AKI was induced, with early post ischemic GFR reductions of approximately 80%. However, renal functional recovery promptly ensued, and over the ensuing 40 weeks, only a 10% GFR reduction (NS) below normal values (in sham operated rats) was observed. Of note, however, is that renal histology demonstrated modest interstitial fibrosis and an approximate 35% reduction in the density of peritubular capillaries. The latter was posited to contribute to ongoing ischemic / hypoxic tissue damage. Thus, while these workers demonstrated that ischemic AKI did, indeed, produce permanent renal damage after 40 weeks, severe progressive CKD did not ensue. Hence, these experimental findings are consistent with the above noted clinical conclusion: that “the vast majority of AKI results in clinically normal renal function despite frequent subclinical and histological defects” (1).

Nath et al have also evaluated the issue of long term AKI outcomes using a model of repetitive nephrotoxic insults (5). These workers employed weekly injections of hypertonic glycerol into the hind limbs of rats, inducing repetitive bouts of muscle necrosis and myohemoglobinuria. After a single dose of glycerol, severe AKI was produced. They then administered glycerol on a weekly basis for 6 months, and at the end of these injections, an approximate 1/3^rd loss of GFR was observed. However, no loss of renal parenchymal mass (renal weight) and only minimal proteinuria developed. Furthermore, it is notable that in that study, the rats continued to lose body weight. Thus, in light of the renal mass preservation, it is not clear whether cachexia, which is known to reduce GFR, contributed to the observed 1/3rd GFR decline. One potential explanation for this seemingly benign renal outcome despite severe repetitive bouts of rhabdomyolysis is that AKI induces a variety of cytoprotective mechanisms (e.g. an up-regulation of heme oxygenase-1; ref. 6), which could serve to mitigate ongoing renal damage. These considerations notwithstanding, the data would certainly seem to challenge the accepted notion that repetitive bouts of AKI culminates in progressive renal disease.

Unilateral ischemic renal injury recapitulates post AKI progression to ESRD

Sixty years ago, Koletsky made the observation that if 60 min of unilateral ischemia is induced, severe and progressive post-ischemic injury results, culminating in end-stage kidney disease (7). This stands in sharp contrast to the results of Basile, discussed above, whereby 60 min of bilateral ischemic injury failed to produce sustained renal failure. Based on these divergent results, we recently tested the hypothesis that unilateral, but not bilateral, ischemic renal injury induces progressive renal disease. To this end, we induced 30 min of unilateral ischemia in the mouse, and severe progressive renal failure ensued (8,9). Thus, within just 2–3 weeks post unilateral ischemia, severe interstitial inflammation, fibrosis, tubular dropout, and a 67% loss of renal mass were observed. Conversely, when bilateral ischemia was induced, renal mass was well preserved. To further illustrate this point, we induced 30 min of left kidney unilateral ischemia + progressive right kidney ischemic damage, ranging from 0–20 min of renal artery occlusion. By so doing, we were able to induce graded azotemia, as assessed 24 hrs post surgery (10). The key finding of that study was that the greater the length of contralateral ischemic damage, the greater was the initial azotemia, and the lesser the amount of left kidney damage, as assessed 2 weeks later. Thus, unilateral ischemia produced end stage kidney disease, but this could be almost completely prevented by inducing contralateral ischemic damage. Stated differently, the greater the degree of initial azotemia (24 hr BUN levels), the lesser the degree of renal parenchymal loss in the 30 min post ischemic kidney (depicted in Figure 1).

Figure 1. The relationship between the severity of initial azotemia and the ultimate loss of post ischemic renal parenchymal mass.

Thirty min of ischemia was imposed on the left kidney with variable degrees of ischemia (0–20 min) imposed on the right kidney. As expected, this produced variable degrees of azotemia, as assessed 24 hrs post surgery. The greater the degree of initial azotemia (assessed 24 hrs post 30 min left ischemia + 0–20 min right ischemia), the lesser the percent loss of left renal mass (assessed by renal weight) 2 weeks post surgery.

The state of uremia can confer cytoprotective effects

The strong correlation between renal protection and the initial severity of azotemia in the above study raised an intriguing hypothesis: that acute uremia can exert cytoprotective and anti-inflammatory effects, and that this “uremic brake” prevents renal disease progression. A series of studies from our laboratory provide direct experimental support for this concept (e.g. ref. 11,12). First, when unilateral ischemia was induced in the presence of uremia (induced by damaging the contralateral kidney), a marked suppression of pro-inflammatory / pro-fibrotic cytokines, and a marked up-regulation of cytoprotective proteins (IL-10 and HO-1), were observed. Hence, the balance between pro-inflammatory vs. anti-inflammatory molecules appeared to be tilted towards the anti-inflammatory state. Second, when mice with severe renal failure underwent peritoneal dialysis, addition of that dialysate to cultured tubular cells conferred a cytoprotective /anti-inflammatory state; Third, when human urine samples were added to freshly isolated mouse proximal tubules (creating experimental in vitro uremia), the tubules became resistant to hypoxic attack; and fourth, the creation of experimental uremia in mice in the presence of normal kidneys (by performing peritoneal dialysis using normal human urine) protected the kidneys from subsequent ischemic damage (13). In composite, these studies provide a potential explanation for why unilateral ischemic injury (i.e., in the absence of uremia), but not bilateral ischemic injury (in the presence of uremia), leads to progressive renal disease. The key question is what component of the uremic milieu is responsible for the observed cytoprotective / anti-inflammatory state. To date, it appears that it is a low molecular weight, heat labile, substance that resists freeze - thawing. Clearly, if this substance (or substances) could be identified, novel prophylactic and therapeutic opportunities might result.

Potential therapeutic strategies for preventing AKI progression to CKD

Although the clinical relevance of the unilateral ischemia model may be questioned (given that AKI is typically a bilateral disease), it does provide a unique opportunity to study pathways of post AKI disease progression. For example, Bonventre’s group (14), studying this model, has proposed that progression is mediated via proximal tubule cell cycle arrest at the G2/M phase, resulting in the generation of profibrotic factors including cytokines, growth factors, and matrix proteins. Conversely, this block was not present with bilateral ischemic damage. Alternatively, we have demonstrated that unilateral renal ischemia leads to a marked remodeling of chromatin at pro-inflammatory genes (9), potentially contributing to unbridled gene activation, and presumably ongoing renal inflammation and fibrosis. The importance of intrarenal inflammation in this model was underscored by our observation that high dose glucocorticoid therapy attenuated (but did not eliminate) progressive post ischemic renal disease (8).

As noted previously, Basile et al (3,4) have provided the novel insight that one potential reason for post ischemic renal disease progression in rodents is the loss of peritubular capillaries, a process that presumably induces ongoing hypoxic tissue damage. These observations led us to question whether ongoing tissue hypoxia might be exacerbated by endothelin 1- mediated renal vasoconstriction. Towards this end, mice were subjected to 30 min of unilateral renal ischemia (15). ET-1 mRNA and protein levels, as well as the mRNAs for ET-1’s A and B receptors (which mediate vasoconstriction and vasodilation, respectively), were assessed 24 hrs or two weeks later. The impacts of potent ETA or ETB receptor- specific antagonists (Atrasentan, BQ-788, respectively) on post-ischemic disease progression were also assessed. Unilateral ischemia caused massive up-regulation of both renal cortical ET-1 mRNA and ET-1 protein levels (~10 and 50 fold respectively). Furthermore, parallel increases in the ETA, but not the ETB, receptor mRNA, was also observed. Chromatin immunoprecipitation assay revealed extensive ‘gene activating’ histone remodeling (methylation, acetylation, H2A.Z enrichment) and increased RNA polymerase II (PolI II) binding at the ET-1 gene, thereby providing a potential explanation for ET-1 over-expression. The functional significance of the ET-1 over-expression was indicated by the observations that the ETA receptor blocker, Atrasentan, administered either pre + post-ischemia, or started 24 hrs post- ischemia, conferred marked protective effects, preserving renal mass. That ETA blockade also lowered renal tissue lactate levels suggested that its protective action was mediated, at least in part, by improvements in renal oxygen delivery. Conversely, ETB receptor blockade was without beneficial effect. Based on these studies, it would appear that ischemic renal injury evokes progressive activation of the ET-1 gene, possibly by induction of ‘gene activating’ histone remodeling at this site. Furthermore, these studies provide the first evidence that ET-1 can play a critical role in post-ischemic AKI progression to CKD, at least as studied in the unilateral ischemia model. Thus, even when AKI initiates renal disease progression, it is clearly not inevitable.

Conclusions

It has long been recognized that AKI has the capacity to initiate CKD. However, the frequency of disease progression remains a subject of considerable debate. Much of the clinical literature would suggest that post AKI renal disease may simply be a result of ‘healing defects’, rather than true progression. The experimental literature would appear to support this conclusion, given that severe bilateral ischemic renal injury or recurrent nephrotoxic AKI has not been demonstrated to lead to severe progressive, or end stage, kidney disease. Multiple factors undoubtedly impact whether or not disease progression results, e.g., whether or not pre-existing renal disease exists, patient age, associated vascular disease, and whether or not recurrent bouts of injury occur. But if disease progression does result, it is clearly not inevitable. For example, selected mechanistic pathways can potentially be blocked (e.g., endothelin 1 overexpression, renal inflammation). Finally, if indeed, uremic cytoprotectant / anti-inflammatory compounds truly exist, and if they can be identified and clinically deployed, then completely new pharmacologic approaches to preventing AKI induction and progression may emerge.